Earth's AISatellite + LiDAR + AI

Satellite + LiDAR + AI

Forest 3D Intelligence with Satellite, LiDAR + AI

Draw one area. LiDAR loads first for the measured canopy surface, AI trains once on that same area, then changing the year reuses the trained model to predict canopy structure from satellite imagery.

One AOITrain once · predict by yearLiDAR and AI compared on the same footprint
LiDAR—
Tree forms*—
Mean— m
Min— m
Max— m
P90— m
AI—
Tree forms*—
Mean— m
Min— m
Max— m
P90— m
MAE— mTraining validation
RMSE— mTraining validation
R²—Training validation
Training year—model trained once
Prediction year—same AOI
LiDAR year—EA VOM reference

*Tree forms are occupied canopy grid cells used for visual comparison, not individual-tree detections.

Interpretation

LiDAR measured. AI modelled.

LiDAR is the measured canopy-height reference for the selected AOI. AI predictions reuse the model trained on that AOI and show how the canopy signal changes when satellite imagery from another year is supplied.

Forest 3D Intelligence FAQ

Frequently asked questions

A quick guide to what Forest 3D Intelligence can show, how often it can be updated and how the results should be used.

What is Earth's AI Forest 3D Intelligence?
Earth's AI Forest 3D Intelligence trains a canopy-height regression model from Sentinel-1 and Sentinel-2 features using Environment Agency LiDAR-derived height labels.
What does Forest 3D Intelligence monitor or analyse?
It estimates landscape-scale canopy height and forest structure, then returns a 2D height layer and a modelled 3D canopy surface.
What data does Forest 3D Intelligence use?
The v1 workflow uses Sentinel-1 GRD, Sentinel-2 SR Harmonized and Environment Agency LiDAR Vegetation Object Model data served through WCS for England.
How often can Forest 3D Intelligence be updated?
A model can be retrained for a selected Sentinel year. Operational refresh frequency depends on Sentinel availability and the suitability of training labels for the target geography.
How should Forest 3D Intelligence results be used?
Use the result as an AI-estimated canopy-height product. It is not a reconstructed LiDAR point cloud, individual-tree survey or substitute for airborne LiDAR where precise 3D geometry is required.

Speak to Earth's AI

Need canopy-height or forest-structure modelling across a larger region?

Earth's AI can support LiDAR/Sentinel training-data pipelines, geographically held-out validation and deployment of custom forest-structure models.

earthsaitech@gmail.com